Added my conclusions for now

Author: TarrySingh

deep-learning/Models_Algorithms_More/Which Optimizer Really Kicks Ass - Adam, SGD, RMSprop or Momentum.ipynb

@@ -1579,6 +1579,24 @@
     "plt.show()"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## My observations:\n",
+    "\n",
+    "- Adam,Nadam have the lowest training loss but you see that SGD & RMSprop come a close second!\n",
+    "- SGD+ Nest performs rather poorly above though!\n",
+    "- Adam still wins hands down in Val training as well\n",
+    "\n",
+    "**Concluding (for now)**\n",
+    "There's stil la lot to explore but we can see that Adam is kicking ass.\n",
+    "\n",
+    "We basically used a simple FFN (Feed forward neural network) with just three hidden layer, next experiment I'll try with 6 hidden layers.\n",
+    "\n",
+    "So, stay tuned!"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,